mirrors/cosmopolitan
1
0
Fork 0
mirror of https://github.com/jart/cosmopolitan.git synced 2024-05-18 03:22:40 +00:00
Code Issues Projects Releases Wiki Activity
cosmopolitan/libc/stdio/iconv.c
Justine Tunney 957c61cbbf
Release Cosmopolitan v3.3 
This change upgrades to GCC 12.3 and GNU binutils 2.42. The GNU linker
appears to have changed things so that only a single de-duplicated str
table is present in the binary, and it gets placed wherever the linker
wants, regardless of what the linker script says. To cope with that we
need to stop using .ident to embed licenses. As such, this change does
significant work to revamp how third party licenses are defined in the
codebase, using `.section .notice,"aR",@progbits`.

This new GCC 12.3 toolchain has support for GNU indirect functions. It
lets us support __target_clones__ for the first time. This is used for
optimizing the performance of libc string functions such as strlen and
friends so far on x86, by ensuring AVX systems favor a second codepath
that uses VEX encoding. It shaves some latency off certain operations.
It's a useful feature to have for scientific computing for the reasons
explained by the test/libcxx/openmp_test.cc example which compiles for
fifteen different microarchitectures. Thanks to the upgrades, it's now
also possible to use newer instruction sets, such as AVX512FP16, VNNI.

Cosmo now uses the %gs register on x86 by default for TLS. Doing it is
helpful for any program that links `cosmo_dlopen()`. Such programs had
to recompile their binaries at startup to change the TLS instructions.
That's not great, since it means every page in the executable needs to
be faulted. The work of rewriting TLS-related x86 opcodes, is moved to
fixupobj.com instead. This is great news for MacOS x86 users, since we
previously needed to morph the binary every time for that platform but
now that's no longer necessary. The only platforms where we need fixup
of TLS x86 opcodes at runtime are now Windows, OpenBSD, and NetBSD. On
Windows we morph TLS to point deeper into the TIB, based on a TlsAlloc
assignment, and on OpenBSD/NetBSD we morph %gs back into %fs since the
kernels do not allow us to specify a value for the %gs register.

OpenBSD users are now required to use APE Loader to run Cosmo binaries
and assimilation is no longer possible. OpenBSD kernel needs to change
to allow programs to specify a value for the %gs register, or it needs
to stop marking executable pages loaded by the kernel as mimmutable().

This release fixes __constructor__, .ctor, .init_array, and lastly the
.preinit_array so they behave the exact same way as glibc.

We no longer use hex constants to define math.h symbols like M_PI.
2024-02-20 13:27:59 -08:00

723 lines
17 KiB
C
Raw Permalink Blame History

/*-*- mode:c;indent-tabs-mode:nil;c-basic-offset:2;tab-width:8;coding:utf-8 -*-│
│ vi: set noet ft=c ts=2 sts=2 sw=2 fenc=utf-8 :vi │
╚──────────────────────────────────────────────────────────────────────────────╝
│ │
│ Musl Libc │
│ Copyright © 2005-2014 Rich Felker, et al. │
│ │
│ Permission is hereby granted, free of charge, to any person obtaining │
│ a copy of this software and associated documentation files (the │
│ "Software"), to deal in the Software without restriction, including │
│ without limitation the rights to use, copy, modify, merge, publish, │
│ distribute, sublicense, and/or sell copies of the Software, and to │
│ permit persons to whom the Software is furnished to do so, subject to │
│ the following conditions: │
│ │
│ The above copyright notice and this permission notice shall be │
│ included in all copies or substantial portions of the Software. │
│ │
│ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, │
│ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF │
│ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. │
│ IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY │
│ CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, │
│ TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE │
│ SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. │
│ │
╚─────────────────────────────────────────────────────────────────────────────*/
#include "libc/stdio/iconv.h"
#include "libc/errno.h"
#include "libc/mem/mem.h"
#include "libc/str/locale.h"
#include "libc/str/str.h"
#include "libc/thread/tls.h"
// clang-format off
__static_yoink("musl_libc_notice");
#define UTF_32BE 0300
#define UTF_16LE 0301
#define UTF_16BE 0302
#define UTF_32LE 0303
#define UCS2BE 0304
#define UCS2LE 0305
#define WCHAR_T 0306
#define US_ASCII 0307
#define UTF_8 0310
#define UTF_16 0312
#define UTF_32 0313
#define UCS2 0314
#define EUC_JP 0320
#define SHIFT_JIS 0321
#define ISO2022_JP 0322
#define GB18030 0330
#define GBK 0331
#define GB2312 0332
#define BIG5 0340
#define EUC_KR 0350
/* Definitions of charmaps. Each charmap consists of:
* 1. Empty-string-terminated list of null-terminated aliases.
* 2. Special type code or number of elided quads of entries.
* 3. Character table (size determined by field 2), consisting
* of 5 bytes for every 4 characters, interpreted as 10-bit
* indices into the legacy_chars table. */
static const unsigned char charmaps[] =
"utf8\0char\0\0\310"
"wchart\0\0\306"
"ucs2be\0\0\304"
"ucs2le\0\0\305"
"utf16be\0\0\302"
"utf16le\0\0\301"
"ucs4be\0utf32be\0\0\300"
"ucs4le\0utf32le\0\0\303"
"ascii\0usascii\0iso646\0iso646us\0\0\307"
"utf16\0\0\312"
"ucs4\0utf32\0\0\313"
"ucs2\0\0\314"
"eucjp\0\0\320"
"shiftjis\0sjis\0\0\321"
"iso2022jp\0\0\322"
"gb18030\0\0\330"
"gbk\0\0\331"
"gb2312\0\0\332"
"big5\0bigfive\0cp950\0big5hkscs\0\0\340"
"euckr\0ksc5601\0ksx1001\0cp949\0\0\350"
#include "libc/stdio/codepages.inc"
;
#pragma GCC diagnostic ignored "-Wmissing-braces"
/* Table of characters that appear in legacy 8-bit codepages,
* limited to 1024 slots (10 bit indices). The first 256 entries
* are elided since those characters are obviously all included. */
static const unsigned short legacy_chars[] = {
#include "libc/stdio/legacychars.inc"
};
static const unsigned short jis0208[84][94] = {
#include "libc/stdio/jis0208.inc"
};
static const unsigned short gb18030[126][190] = {
#include "libc/stdio/gb18030.inc"
};
static const unsigned short big5[89][157] = {
#include "libc/stdio/big5.inc"
};
static const unsigned short hkscs[] = {
#include "libc/stdio/hkscs.inc"
};
static const unsigned short ksc[93][94] = {
#include "libc/stdio/ksc.inc"
};
static const unsigned short rev_jis[] = {
#include "libc/stdio/revjis.inc"
};
static int fuzzycmp(const unsigned char *a, const unsigned char *b)
{
for (; *a && *b; a++, b++) {
while (*a && (*a|32U)-'a'>26 && *a-'0'>10U) a++;
if ((*a|32U) != *b) return 1;
}
return *a != *b;
}
static size_t find_charmap(const void *name)
{
const unsigned char *s;
if (!*(char *)name) name=charmaps; /* "utf8" */
for (s=charmaps; *s; ) {
if (!fuzzycmp(name, s)) {
for (; *s; s+=strlen((void *)s)+1);
return s+1-charmaps;
}
s += strlen((void *)s)+1;
if (!*s) {
if (s[1] > 0200) s+=2;
else s+=2+(64U-s[1])*5;
}
}
return -1;
}
struct stateful_cd {
iconv_t base_cd;
unsigned state;
};
static iconv_t combine_to_from(size_t t, size_t f)
{
return (void *)(f<<16 | t<<1 | 1);
}
static size_t extract_from(iconv_t cd)
{
return (size_t)cd >> 16;
}
static size_t extract_to(iconv_t cd)
{
return (size_t)cd >> 1 & 0x7fff;
}
iconv_t iconv_open(const char *to, const char *from)
{
size_t f, t;
struct stateful_cd *scd;
if ((t = find_charmap(to))==-1
|| (f = find_charmap(from))==-1
|| (charmaps[t] >= 0330)) {
errno = EINVAL;
return (iconv_t)-1;
}
iconv_t cd = combine_to_from(t, f);
switch (charmaps[f]) {
case UTF_16:
case UTF_32:
case UCS2:
case ISO2022_JP:
scd = malloc(sizeof *scd);
if (!scd) return (iconv_t)-1;
scd->base_cd = cd;
scd->state = 0;
cd = (iconv_t)scd;
}
return cd;
}
int iconv_close(iconv_t cd)
{
if (!((size_t)cd & 1)) free((void *)cd);
return 0;
}
static unsigned get_16(const unsigned char *s, int e)
{
e &= 1;
return s[e]<<8 | s[1-e];
}
static void put_16(unsigned char *s, unsigned c, int e)
{
e &= 1;
s[e] = c>>8;
s[1-e] = c;
}
static unsigned get_32(const unsigned char *s, int e)
{
e &= 3;
return (s[e]+0U)<<24 | s[e^1]<<16 | s[e^2]<<8 | s[e^3];
}
static void put_32(unsigned char *s, unsigned c, int e)
{
e &= 3;
s[e^0] = c>>24;
s[e^1] = c>>16;
s[e^2] = c>>8;
s[e^3] = c;
}
/* Adapt as needed */
#define mbrtowc_utf8 mbrtowc
#define wctomb_utf8 wctomb
static unsigned legacy_map(const unsigned char *map, unsigned c)
{
if (c < 4*map[-1]) return c;
unsigned x = c - 4*map[-1];
x = (map[x*5/4]>>(2*x%8)) | ((map[x*5/4+1]<<(8-(2*x%8))) & 1023);
return x < 256 ? x : legacy_chars[x-256];
}
static unsigned uni_to_jis(unsigned c)
{
unsigned nel = sizeof rev_jis / sizeof *rev_jis;
unsigned d, j, i, b = 0;
for (;;) {
i = nel/2;
j = rev_jis[b+i];
d = jis0208[j/256][j%256];
if (d==c) return j + 0x2121;
else if (nel == 1) return 0;
else if (c < d)
nel /= 2;
else {
b += i;
nel -= nel/2;
}
}
}
size_t iconv(iconv_t cd, char **restrict in, size_t *restrict inb, char **restrict out, size_t *restrict outb)
{
size_t x=0;
struct stateful_cd *scd=0;
if (!((size_t)cd & 1)) {
scd = (void *)cd;
cd = scd->base_cd;
}
unsigned to = extract_to(cd);
unsigned from = extract_from(cd);
const unsigned char *map = charmaps+from+1;
const unsigned char *tomap = charmaps+to+1;
mbstate_t st = {0};
wchar_t wc;
unsigned c, d;
size_t k, l;
int err;
unsigned char type = map[-1];
unsigned char totype = tomap[-1];
locale_t *ploc = (locale_t *)&__get_tls()->tib_locale;
locale_t loc = *ploc;
if (!in || !*in || !*inb) return 0;
*ploc = 0; // TODO(jart): UTF8_LOCALE?
for (; *inb; *in+=l, *inb-=l) {
c = *(unsigned char *)*in;
l = 1;
switch (type) {
case UTF_8:
if (c < 128) break;
l = mbrtowc_utf8(&wc, *in, *inb, &st);
if (l == (size_t)-1) goto ilseq;
if (l == (size_t)-2) goto starved;
c = wc;
break;
case US_ASCII:
if (c >= 128) goto ilseq;
break;
case WCHAR_T:
l = sizeof(wchar_t);
if (*inb < l) goto starved;
c = *(wchar_t *)*in;
if (0) {
case UTF_32BE:
case UTF_32LE:
l = 4;
if (*inb < 4) goto starved;
c = get_32((void *)*in, type);
}
if (c-0xd800u < 0x800u || c >= 0x110000u) goto ilseq;
break;
case UCS2BE:
case UCS2LE:
case UTF_16BE:
case UTF_16LE:
l = 2;
if (*inb < 2) goto starved;
c = get_16((void *)*in, type);
if ((unsigned)(c-0xdc00) < 0x400) goto ilseq;
if ((unsigned)(c-0xd800) < 0x400) {
if (type-UCS2BE < 2U) goto ilseq;
l = 4;
if (*inb < 4) goto starved;
d = get_16((void *)(*in + 2), type);
if ((unsigned)(d-0xdc00) >= 0x400) goto ilseq;
c = ((c-0xd7c0)<<10) + (d-0xdc00);
}
break;
case UCS2:
case UTF_16:
l = 0;
if (!scd->state) {
if (*inb < 2) goto starved;
c = get_16((void *)*in, 0);
scd->state = type==UCS2
? c==0xfffe ? UCS2LE : UCS2BE
: c==0xfffe ? UTF_16LE : UTF_16BE;
if (c == 0xfffe || c == 0xfeff)
l = 2;
}
type = scd->state;
continue;
case UTF_32:
l = 0;
if (!scd->state) {
if (*inb < 4) goto starved;
c = get_32((void *)*in, 0);
scd->state = c==0xfffe0000 ? UTF_32LE : UTF_32BE;
if (c == 0xfffe0000 || c == 0xfeff)
l = 4;
}
type = scd->state;
continue;
case SHIFT_JIS:
if (c < 128) break;
if (c-0xa1 <= 0xdf-0xa1) {
c += 0xff61-0xa1;
break;
}
l = 2;
if (*inb < 2) goto starved;
d = *((unsigned char *)*in + 1);
if (c-129 <= 159-129) c -= 129;
else if (c-224 <= 239-224) c -= 193;
else goto ilseq;
c *= 2;
if (d-64 <= 158-64) {
if (d==127) goto ilseq;
if (d>127) d--;
d -= 64;
} else if (d-159 <= 252-159) {
c++;
d -= 159;
}
c = jis0208[c][d];
if (!c) goto ilseq;
break;
case EUC_JP:
if (c < 128) break;
l = 2;
if (*inb < 2) goto starved;
d = *((unsigned char *)*in + 1);
if (c==0x8e) {
c = d;
if (c-0xa1 > 0xdf-0xa1) goto ilseq;
c += 0xff61 - 0xa1;
break;
}
c -= 0xa1;
d -= 0xa1;
if (c >= 84 || d >= 94) goto ilseq;
c = jis0208[c][d];
if (!c) goto ilseq;
break;
case ISO2022_JP:
if (c >= 128) goto ilseq;
if (c == '\033') {
l = 3;
if (*inb < 3) goto starved;
c = *((unsigned char *)*in + 1);
d = *((unsigned char *)*in + 2);
if (c != '(' && c != '$') goto ilseq;
switch (128*(c=='$') + d) {
case 'B': scd->state=0; continue;
case 'J': scd->state=1; continue;
case 'I': scd->state=4; continue;
case 128+'@': scd->state=2; continue;
case 128+'B': scd->state=3; continue;
}
goto ilseq;
}
switch (scd->state) {
case 1:
if (c=='\\') c = 0xa5;
if (c=='~') c = 0x203e;
break;
case 2:
case 3:
l = 2;
if (*inb < 2) goto starved;
d = *((unsigned char *)*in + 1);
c -= 0x21;
d -= 0x21;
if (c >= 84 || d >= 94) goto ilseq;
c = jis0208[c][d];
if (!c) goto ilseq;
break;
case 4:
if (c-0x60 < 0x1f) goto ilseq;
if (c-0x21 < 0x5e) c += 0xff61-0x21;
break;
}
break;
case GB2312:
if (c < 128) break;
if (c < 0xa1) goto ilseq;
case GBK:
case GB18030:
if (c < 128) break;
c -= 0x81;
if (c >= 126) goto ilseq;
l = 2;
if (*inb < 2) goto starved;
d = *((unsigned char *)*in + 1);
if (d < 0xa1 && type == GB2312) goto ilseq;
if (d-0x40>=191 || d==127) {
if (d-'0'>9 || type != GB18030)
goto ilseq;
l = 4;
if (*inb < 4) goto starved;
c = (10*c + d-'0') * 1260;
d = *((unsigned char *)*in + 2);
if (d-0x81>126) goto ilseq;
c += 10*(d-0x81);
d = *((unsigned char *)*in + 3);
if (d-'0'>9) goto ilseq;
c += d-'0';
c += 128;
for (d=0; d<=c; ) {
k = 0;
for (int i=0; i<126; i++)
for (int j=0; j<190; j++)
if (gb18030[i][j]-d <= c-d)
k++;
d = c+1;
c += k;
}
break;
}
d -= 0x40;
if (d>63) d--;
c = gb18030[c][d];
break;
case BIG5:
if (c < 128) break;
l = 2;
if (*inb < 2) goto starved;
d = *((unsigned char *)*in + 1);
if (d-0x40>=0xff-0x40 || d-0x7f<0xa1-0x7f) goto ilseq;
d -= 0x40;
if (d > 0x3e) d -= 0x22;
if (c-0xa1>=0xfa-0xa1) {
if (c-0x87>=0xff-0x87) goto ilseq;
if (c < 0xa1) c -= 0x87;
else c -= 0x87 + (0xfa-0xa1);
c = (hkscs[4867+(c*157+d)/16]>>(c*157+d)%16)%2<<17
| hkscs[c*157+d];
/* A few HKSCS characters map to pairs of UCS
* characters. These are mapped to surrogate
* range in the hkscs table then hard-coded
* here. Ugly, yes. */
if (c/256 == 0xdc) {
union {
char c[8];
wchar_t wc[2];
} tmp;
char *ptmp = tmp.c;
size_t tmpx = iconv(combine_to_from(to, find_charmap("utf8")),
&(char *){"\303\212\314\204"
"\303\212\314\214"
"\303\252\314\204"
"\303\252\314\214"
+c%256}, &(size_t){4},
&ptmp, &(size_t){sizeof tmp});
size_t tmplen = ptmp - tmp.c;
if (tmplen > *outb) goto toobig;
if (tmpx) x++;
memcpy(*out, &tmp, tmplen);
*out += tmplen;
*outb -= tmplen;
continue;
}
if (!c) goto ilseq;
break;
}
c -= 0xa1;
c = big5[c][d]|(c==0x27&&(d==0x3a||d==0x3c||d==0x42))<<17;
if (!c) goto ilseq;
break;
case EUC_KR:
if (c < 128) break;
l = 2;
if (*inb < 2) goto starved;
d = *((unsigned char *)*in + 1);
c -= 0xa1;
d -= 0xa1;
if (c >= 93 || d >= 94) {
c += (0xa1-0x81);
d += 0xa1;
if (c >= 93 || ((c>=0xc6-0x81) && d>0x52))
goto ilseq;
if (d-'A'<26) d = d-'A';
else if (d-'a'<26) d = d-'a'+26;
else if (d-0x81<0xff-0x81) d = d-0x81+52;
else goto ilseq;
if (c < 0x20) c = 178*c + d;
else c = 178*0x20 + 84*(c-0x20) + d;
c += 0xac00;
for (d=0xac00; d<=c; ) {
k = 0;
for (int i=0; i<93; i++)
for (int j=0; j<94; j++)
if (ksc[i][j]-d <= c-d)
k++;
d = c+1;
c += k;
}
break;
}
c = ksc[c][d];
if (!c) goto ilseq;
break;
default:
if (!c) break;
c = legacy_map(map, c);
if (!c) goto ilseq;
}
switch (totype) {
case WCHAR_T:
if (*outb < sizeof(wchar_t)) goto toobig;
*(wchar_t *)*out = c;
*out += sizeof(wchar_t);
*outb -= sizeof(wchar_t);
break;
case UTF_8:
if (*outb < 4) {
char tmp[4];
k = wctomb_utf8(tmp, c);
if (*outb < k) goto toobig;
memcpy(*out, tmp, k);
} else k = wctomb_utf8(*out, c);
*out += k;
*outb -= k;
break;
case US_ASCII:
if (c > 0x7f) subst: x++, c='*';
default:
if (*outb < 1) goto toobig;
if (c<256 && c==legacy_map(tomap, c)) {
revout:
if (*outb < 1) goto toobig;
*(*out)++ = c;
*outb -= 1;
break;
}
d = c;
for (c=4*totype; c<256; c++) {
if (d == legacy_map(tomap, c)) {
goto revout;
}
}
goto subst;
case SHIFT_JIS:
if (c < 128) goto revout;
if (c == 0xa5) {
x++;
c = '\\';
goto revout;
}
if (c == 0x203e) {
x++;
c = '~';
goto revout;
}
if (c-0xff61 <= 0xdf-0xa1) {
c += 0xa1 - 0xff61;
goto revout;
}
c = uni_to_jis(c);
if (!c) goto subst;
if (*outb < 2) goto toobig;
d = c%256;
c = c/256;
*(*out)++ = (c+1)/2 + (c<95 ? 112 : 176);
*(*out)++ = c%2 ? d + 31 + d/96 : d + 126;
*outb -= 2;
break;
case EUC_JP:
if (c < 128) goto revout;
if (c-0xff61 <= 0xdf-0xa1) {
c += 0x0e00 + 0x21 - 0xff61;
} else {
c = uni_to_jis(c);
}
if (!c) goto subst;
if (*outb < 2) goto toobig;
*(*out)++ = c/256 + 0x80;
*(*out)++ = c%256 + 0x80;
*outb -= 2;
break;
case ISO2022_JP:
if (c < 128) goto revout;
if (c-0xff61 <= 0xdf-0xa1 || c==0xa5 || c==0x203e) {
if (*outb < 7) goto toobig;
*(*out)++ = '\033';
*(*out)++ = '(';
if (c==0xa5) {
*(*out)++ = 'J';
*(*out)++ = '\\';
} else if (c==0x203e) {
*(*out)++ = 'J';
*(*out)++ = '~';
} else {
*(*out)++ = 'I';
*(*out)++ = c-0xff61+0x21;
}
*(*out)++ = '\033';
*(*out)++ = '(';
*(*out)++ = 'B';
*outb -= 7;
break;
}
c = uni_to_jis(c);
if (!c) goto subst;
if (*outb < 8) goto toobig;
*(*out)++ = '\033';
*(*out)++ = '$';
*(*out)++ = 'B';
*(*out)++ = c/256;
*(*out)++ = c%256;
*(*out)++ = '\033';
*(*out)++ = '(';
*(*out)++ = 'B';
*outb -= 8;
break;
case UCS2:
totype = UCS2BE;
case UCS2BE:
case UCS2LE:
case UTF_16:
case UTF_16BE:
case UTF_16LE:
if (c < 0x10000 || totype-UCS2BE < 2U) {
if (c >= 0x10000) c = 0xFFFD;
if (*outb < 2) goto toobig;
put_16((void *)*out, c, totype);
*out += 2;
*outb -= 2;
break;
}
if (*outb < 4) goto toobig;
c -= 0x10000;
put_16((void *)*out, (c>>10)|0xd800, totype);
put_16((void *)(*out + 2), (c&0x3ff)|0xdc00, totype);
*out += 4;
*outb -= 4;
break;
case UTF_32:
totype = UTF_32BE;
case UTF_32BE:
case UTF_32LE:
if (*outb < 4) goto toobig;
put_32((void *)*out, c, totype);
*out += 4;
*outb -= 4;
break;
}
}
*ploc = loc;
return x;
ilseq:
err = EILSEQ;
x = -1;
goto end;
toobig:
err = E2BIG;
x = -1;
goto end;
starved:
err = EINVAL;
x = -1;
end:
errno = err;
*ploc = loc;
return x;
}
Reference in a new issue View git blame Copy permalink